Create qm sanity tests

[pbrilla-rh]

We have several test cases for FFI, but we are missing some basic/simple sanity tests for qm, if it is installed properly and/or if environment started successfully

Adding Note: I would add updating tiear-0 test to verify through bluchictl

If there are 2 bluechi nodes connected to controller from host and QM it means that QM service is up and podman container is running

[dougsland]

@pengshanyu are you interested in this one? I am sure @pbrilla-rh and I can share more info if needed. Assigning to you for now.

[Yarboa]

@pengshanyu

I assume tier-0 tests which run on different scenario in the pipeline sould run also with the filter: tag:ffi of /plans/e2e/ffi

Please use this test, tests/qm-unit-files/main.fmf

that run in different test setups , you can manipulate the same
like here /tests/qm-connectivity/tier0 /tests/qm-connectivity/tier1

Add ffi tag for test /tests/qm-connectivity/ffi Whenever FFI filter is selected, qm-connectivity, with some ENV changes will verify ffi setup

Consider also change the tags, since now test tier-0 will be ffi,

[pengshanyu]

@Yarboa Thank you for sharing.

[pengshanyu]

Hi @Yarboa, it seems that the tier0 test mentioned by Pavol is different from the one you mentioned in the comment above. What test cases do we want tier0 to include? Thanks.

[Yarboa]

Correct this repo evolved within time Could be that tier-0 tests running in that gate, should be renamed, I think it is not so important for now that we have the FFI test environemt

But, I could think of tier-0 tests as per that one:

1. Verify qm is up and running systemctl is-active qm
2. Verify bluechi nodes are connected bluechictl status, 2 nodes connected
3. podman run and exec container inside qm with service file
4. /var partition exist

[pengshanyu]

Hi @Yarboa , I setup the qm environment by running set-ffi-env-e2e, and then do the checks mentioned in https://github.com/containers/qm/issues/342#issuecomment-2200525139. Is this correct, or do I need to setup the qm environment in some other way?

I found inside qm, bluechi-agent can not connect to bluechi-controller.

[root@ibm-p8-kvm-03-guest-02 ~]# podman exec -it qm bash -c "systemctl status bluechi-agent" | tail -2
Jul 09 06:10:33 d32c3acf027c bluechi-agent[35]: Connecting to controller on tcp:host=127.0.0.1,port=842
Jul 09 06:10:33 d32c3acf027c bluechi-agent[35]: Registering as 'qm.localrootfs' failed: Transport endpoint is not connected

I know this is because ControllerHost is missing in /etc/qm/bluechi/agent.conf.d/agent.conf but actually, I didn't find the qm bluechi-agent configuration in set-ffi-env-e2e. Can I insert ControllerHost into /etc/qm/bluechi/agent.conf.d/agent.conf like I did with agent-flood to connect qm bluechi-agent and bluechi-controller?

[pengshanyu]

@Yarboa @pbrilla-rh I named the sanity test to tier-0, but there are also tier-0 tests in qm-connectivity and qm-unit-files, how should they be renamed or do we want them to be tier-0 along with the sanity tests?